Neurocognitive and mental health outcomes and association with quality of life among adults living with HIV: a cross-sectional focus on a low-literacy population from coastal Kenya

Objectives Our aim was to compare the neurocognitive performance and mental health outcome of adults living with HIV on antiretroviral therapy with that of community controls, all of low literacy. Furthermore, we also wanted to explore the relationship of these outcomes with quality of life among adults living with HIV. Study design This was a descriptive cross-sectional study. Setting The study was conducted in Kilifi County, a region located at the Kenyan coast. Participants The participants consisted of a consecutive sample of 84 adults living with HIV and 83 randomly selected community controls all with ≤8 years of schooling. All participants were assessed for non-verbal intelligence, verbal working memory and executive functioning. The Major Depression Inventory and a quality of life measure (RAND SF-36) were also administered. Results Using analysis of covariance, we found no statistically significant group differences between adults living with HIV and community controls in all the neurocognitive tests except for a marginal difference in the non-verbal intelligence test (F (1, 158)=3.83, p=0.05). However, depressive scores of adults living with HIV were significantly higher than those of controls (F (1, 158)=11.56, p<0.01). Also, quality of life scores of adults living with HIV were significantly lower than those of controls (F (1, 158)=4.62, p=0.03). For the HIV-infected group, results from multivariable linear regression analysis showed that increasing depressive scores were significantly associated with poorer quality of life (β=−1.17, 95% CI −1.55 to –0.80; p<0.01). Conclusion Our findings suggest that adults of low-literacy levels living with HIV and on antiretroviral medication at the Kenyan coast do not have significant cognitive deficits compared with their uninfected counterparts. However, their mental health, compared with that of HIV-uninfected adults, remains poorer and their quality of life may deteriorate when HIV and depressive symptoms co-occur.


GENERAL COMMENTS
This is a cross-sectional study comparing cognitive and mental health outcomes amongst people living with HIV and HIV-negative controls in Kenya. The article is well written and presents interesting findings that are appropriately discussed within the study's stated limitations. The background focuses nicely on the study's topic, with recent and relevant references throughout. The methods are clearly described and could be replicated, with especially strong detail regarding the measures. The statistics are well done and presented in an accessible manner. The discussion connects strongly with the background and results. Limitations are clearly stated and I appreciated the detail regarding whether community controls could potentially have been HIV-positive. Funding, ethical considerations, and STROBE checklist items are all transparently reported. Overall, a very interesting article that is presented well and makes a contribution to the field. 3. In addition to the Digit Span total score, please also report the mean number of digits that participants managed.

REVIEWER
4. Is there a reference for the parent study? If so, please include it.

Regarding
Limitations, please comment on the appropriateness of the neuropsychological tests used in this low level of education population.

Reviewer #1
1. This is a cross-sectional study comparing cognitive and mental health outcomes amongst people living with HIV and HIV-negative controls in Kenya. The article is well written and presents interesting findings that are appropriately discussed within the study's stated limitations. The background focuses nicely on the study's topic, with recent and relevant references throughout. The methods are clearly described and could be replicated, with especially strong detail regarding the measures. The statistics are well done and presented in an accessible manner. The discussion connects strongly with the background and results. Limitations are clearly stated and I appreciated the detail regarding whether community controls could potentially have been HIV-positive. Funding, ethical considerations, and STROBE checklist items are all transparently reported. Overall, a very interesting article that is presented well and makes a contribution to the field.
We would like to thank the reviewer for the positive appraisal of this work.

I do not see what reporting data on religion contributes to this study. Consider removing it.
We would like to thank the reviewer for this suggestion. We have now removed data reporting about religion.

It is not clear what is meant by Digit Span Highest level reached. Please clarify.
As a clarification, digit span highest level reached here referred to the highest set of digit length that participants managed to reach. The backward digit span was administered under 8 sets of digits of a given series length. Sets 1, 2, and 3 consisted of a series of 3 random digits between 1 and 10 (set 1 had 2 practice series of digits each with a digit length of 2). Sets 4, 5, 6, 7, and 8 also consisted of random digits between 1 to 10, each set having a series of digits of a length corresponding to the set, e.g. set 4 consisted of a series of digits with a digit length of 4. The computed mean was that of the highest set of digits reached by our participants.
To make this clearer, we now refer to this as "Highest set of digits reached" and elaborate what this means in the table legend (Table 2).
3. In addition to the Digit Span total score, please also report the mean number of digits that participants managed.
We computed two things from the backward digit span: 1) the mean total correct score; and 2) the mean of what we now refer to as the "highest set of digits reached". The latter, in our opinion addresses this comment, since the test was administered in sets of digits with increasing level of length.
As explained in comment 2 above, the computed mean of the "highest set of digits reached" reflects the average level, in terms of increasing digit length set, that participants managed.
Currently, we do not have a published reference for the parent study as analysis and write-up is still ongoing.

Regarding Limitations, please comment on the appropriateness of the neuropsychological tests used in this low level of education population.
We would like to thank the reviewer for this comment. We agree that education level may have a partial role on neurocognitive test performance where more educated subjects are likely to perform better than the less educated due to differences in intellectual functioning. We recruited a low literacy sample (both adults living with HIV and community controls). As a result of this bias in selection, their cognitive performance on average, is expected to be lower when compared to populations with varying levels of educational achievement. Even though this bias applies equally for both study groups, the potential lack of familiarity with neurocognitive test requirement/demand restricts the generalization of our findings.
In the limitations section of our manuscript (page 15), we now include the following statement: "Lastly, we recruited and administered neurocognitive tests to a low literacy sample. The potential lack of familiarity with neurocognitive test requirement/demands by our study participants, because of this biased selection, limits the generalizability of our finding. Our findings should be interpreted in the context of a low-literacy population."